Log message #4186792

# At Username Text
# May 17th, 14:06 nils Haha, I don't envy you!
# May 17th, 14:06 neothermic Welcome to my hell, except this codebase started late 2010 on cake1 :)
# May 17th, 14:05 nils I guess. I'm the only developer in it. The project started 2013 so I get quite frustrated with the 2013-version of me. :slightly_smiling_face:
# May 17th, 14:02 neothermic how large the codebase is, how many developers you have, how many cake2-isims you've got, how recent your cake2 install is, how much you rely on removed functionality, etc
# May 17th, 14:01 neothermic I suspect the time it takes varies on multiple factors
# May 17th, 14:01 neothermic I'm slowly doing that now
# May 17th, 14:01 nils Did anyone migrate from 2.x to 3.x? How time consuming should I expect it to be?
# May 17th, 14:00 neothermic but yeah, any thoughts on why this isn't behaving the same?
# May 17th, 13:57 neothermic :)
# May 17th, 13:57 dereuromark I know I am
# May 17th, 13:56 dereuromark you know, there are things in life where you want to be replaced by machines
# May 17th, 13:56 neothermic but this one is odd, as I'm not sure why I'm losing the subkey; is that expected?
# May 17th, 13:56 neothermic well, the basic testcase is me running the change manually, comparing the outputs :)
# May 17th, 13:56 dereuromark always good to have basic test cases around your internal business logic to catch this early on when migrating
# May 17th, 13:55 neothermic but Hash::Extract with the path of `'Report'` returns the data without the 'Report' subkeying
# May 17th, 13:54 neothermic So we have an array that is basically from a find, that contains a key called 'Report', and then has subitems, and each item is an array of id, etc. `Set::extract('/Report'` on the data returns structure like: ```array( (int) 0 => array( 'Report' => array( 'id' => '45'
# May 17th, 13:52 neothermic argh, Hash::extract is doing something slightly different from Set::extract :O
# May 17th, 13:20 neon1024 Sounds good, sign me up @admad
# May 17th, 13:20 admad are PHP "influencers" available for hire? we could then spend 1 year's cakefest budget on them instead :P
# May 17th, 13:18 np :wave: I forwarded your response to my colleague. Thanks
# May 17th, 13:18 dereuromark One reddit article after 4.0 release might do the trick :P and noise
# May 17th, 13:17 admad I hope people care enough about clean code and it gets us more users :slightly_smiling_face:
# May 17th, 13:15 dereuromark I am shocked and suprised at the same time that both laravel and symfony are not using phpstan/psalm and dismiss any addition as "current code is fine", even though the people showed them that the opposite is true.. Funny that cake already has - and soon will even more so have - the cleanest framework code by far :slightly_smiling_face:
# May 17th, 13:13 nils @dereuromark trying to convince the client to move to 3.x. But it's a huge job, I guess.
# May 17th, 13:10 dereuromark that is too long ago, I cant recall, but could very well be not in schema. You can either PR it, or we ignore it (as I would)
# May 17th, 13:09 dereuromark sry, my brain starts to ignore 2.x :P
# May 17th, 13:08 nils That's only for 3.x, or am I mistaken? I'm running 2.x.
# May 17th, 13:03 dereuromark but it is all documented even in mine: https://github.com/dereuromark/cakephp-queue/tree/master/docs#installation
# May 17th, 13:03 dereuromark make sure you ran the right mgiraitons, that is how it its done as per migrations plugin
# May 17th, 13:00 nils @dereuromark oh, and one more thing. I got an error about a column missing called "priority", it wasn't added when running "Schema create -p Queue" - so I manually added it. Is that a bug maybe?
# May 17th, 12:48 nils The overlap with CronJobs are really smart though, I hit myself for not coming up with that as a solution myself.
# May 17th, 12:48 nils Like built in? I've looked at some third-party ones, but none that really came to my liking. So it all fell on Gearman, which has been working ok. However, lately I've been getting some "Error while sending QUERY packet" - and I do believe it has to do with Gearman
# May 17th, 12:46 dereuromark I asked a while ago if some costs/weight system exists yet for queue jobs, as we run very expensive tasks. But so far I dont think there is something here yet, i was planning to build that in soon as well, as opt-in feature.
# May 17th, 12:45 dereuromark jep
# May 17th, 12:43 nils @dereuromark that's awesome. One more question, all the config-times are put in seconds, right?
# May 17th, 12:42 dereuromark We are running it for our complete release process with tons of jobs per day, works quite well if you have enough server power
# May 17th, 12:40 nils @dereuromark just figured it out on my local machine - it runs fine but this is better. Changing this from Gearman which isn't that stable on my server, can't tell why.
# May 17th, 12:38 dereuromark that is the most important part of the whole system :slightly_smiling_face: otherwise the workers wont run, yeah
# May 17th, 12:38 nils @dereuromark yeah, read it. Maybe the cronjob thing will do the trick?
# May 17th, 12:34 dereuromark if you follow the docs, this is exactly how it works.
# May 17th, 12:30 nils @dereuromark I'm implementing the Queue-plugin you created. I'm still on 2.x though. When implementing it on the production server, is there a way to run the worker in the background? Like a daemon?